Modification: Notary selection process - selecting notaries based on number per region or DataCap required per region

[dkkapur]

Issue Description

In the Governance Call on 2021-04-13, it was raised that having a target number of Notaries per region may not be the most effective form of ensuring good DataCap coverage for clients. We may instead want to discuss the alternative of meeting DataCap projections per region.

Impact

In the previous election cycle, EU ran out of available DataCap a lot sooner than the other regions. Though this was circumstantial, and Notaries from other regions were able to cover - it does raise the flag that Notary DataCap grants for allocations may need to be better aligned with demand in the individual region.

Proposed Solution(s)

TBD

Related Issues

[dkkapur]

In the initial round of Notary elections, < 2 PiB of DataCap was distributed, of which ~75 applications were granted in the Client Onboarding repo. From these, here are some rough notes (please excuse me for errors in this manual tracking of the data):

Region	Number of applications	DataCap allocated (in TiB)
Africa	0	0
Europe	16	75.1
Greater China Region	38	270.36
North America	16	77.2
Oceania	0	0
Asia - GCN	4	8.12
South America	0	0

Based on this - of the ~430.8 TiB across 74 applications:

Region	% of applications	% of DataCap allocated
Europe	22%	17%
Greater China Region	51%	63%
North America	22%	18%
Asia - GCN	5%	2%

However - please note that this is not necessarily a great indicator of demand for DataCap since it only reflects the applications that were accepted rather than the total requests that were made. Additionally - we saw clients apply to notaries in other regions as well, which in many cases yielded a successful allocation.

From the first round of notary grants (1,954 TiB):

Region	DataCap granted to notaries (in TiB)	% of DataCap granted
Europe	210	11%
Greater China Region	1,324	67.5%
North America	400	20.5%
Asia - GCN	20	1%

There's roughly about 1.4 PiB of DataCap still remaining. If we were to increase the circulating DataCap supply substantively, we should consider granting at least the amount that was granted before = >2 PiB. (Please note that #94 is also in active development and there should be a path here for large datasets to be served in the next few weeks.)

If we target total DataCap created (granted/allocated/in deals) = 5 PiB:

Region	Target % total DataCap granted	Target DataCap (in TiB)	Current remaining DataCap	Allocated in this round of elections
Europe	20%	1024	8.52	~1016 TiB
Greater China Region	40%	2048	999	~1049 TiB
North America	20%	1024	234	~790 TiB
Asia - GCN	10%	512	8	~504 TiB
Africa	0-10%	< 512 TiB	0	< 512 TiB
Oceania	0-10%	< 512 TiB	0	< 512 TiB
South America	0-10%	< 512 TiB	0	< 512 TiB

Variance for applications across regions is high - 31 unique applications were filed for the last election, of which 14 were selected with an attempt to get 3 notaries in each region. If we see a lot of new notaries coming in, this could mean ~5-10 notaries in regions like Greater China, EU, NA. The Fil+ ecosystem is still relatively young, and time-to-datacap is high, having more notaries will generally improve maturing the program as well as delivering value through Filecoin faster.

Based on this, proposing that we do the following:

• Increase target notary count per region to 5 for EU, NA, GCR (SA, Asia-GCN, Oceania, South America stay at 3)
• Set max notary count for any region at 7
• Select number of notaries per region based on meeting target notary count, and if that does not bring the region past the expected allocation DataCap from this round (last column in the table below), add notaries till we reach 7.

[KodaRobotDog]

I totally agree with this new strategy.

[dkkapur]

Closing this out for now, will revisit for next election cycle.